Skip to content

Free tool

llms.txt Generator & Validator

Draft an llms.txt from your site’s own pages, edit it, and download the file. Or validate a live llms.txt against the llmstxt.org format.

Or try

The file

What is llms.txt?

llms.txt is a plain text file served at a website’s root that gives AI assistants a curated index of the site: the site’s name, a 1-sentence summary, and short markdown lists of the pages that matter most. It was proposed at llmstxt.org as a convention, the way robots.txt and sitemap.xml began.

A complete example

# Ben's Coffee

> Specialty coffee in Brooklyn, beans shipped nationwide.

## Pages

- [Home](https://benscoffee.com/)
- [Shop](https://benscoffee.com/shop): whole beans and subscriptions
- [About](https://benscoffee.com/about)

## Optional

- [Press](https://benscoffee.com/press)

That is the whole format: 1 H1 title, a blockquote summary, and ## sections of links. The Optional section marks links a reader may skip. The full story of the file, including the evidence on whether engines read it, is in our guide to what llms.txt is.

The checks

What the validator checks

8 rules from the llmstxt.org format. 2 are required, 4 are recommended, and 2 are reported as facts with no judgment attached. The generator’s drafts pass all 8.

Starts with an H1 title

Required

The first line names the site, like # Your site name. It is the only element the format requires, and a file without it fails.

Exactly 1 H1

Required

The title is the file's only H1. A second one breaks the structure a parser expects.

Blockquote summary

Recommended

1 line starting with > directly under the title, saying what the site is. The spec recommends it; a missing summary is a note, not a failure.

Sections use H2 headings

Recommended

Link sections are delimited by ## headings. Deeper headings sit outside the format, so strict parsers may ignore them.

Link lines follow the format

Recommended

Each entry reads - [name](url), with an optional : note after it. A bare URL loses its name to a spec parser.

Link URLs are absolute

Recommended

Full http or https addresses, so a machine reader needs no base URL to resolve them.

Link inventory

Reported

How many links across how many sections. A count, never a judgment: a short file is allowed.

Optional section

Reported

Whether an ## Optional section exists. By convention it holds secondary links a reader may skip when it needs a shorter context.

How it works

How to create an llms.txt file

  1. Enter your address

    Cruelx reads your homepage's title, description and heading, plus the top pages of your sitemap. Nothing is invented: every line comes from what your site already publishes.

  2. Edit the draft

    The draft lands in an editable box and re-checks itself against the format rules as you type. Rewrite the summary, add or drop links, rename anything.

  3. Copy, download, upload

    Copy the text or download llms.txt, then serve it at your site's root. Run the validator afterwards to confirm the live file reads back correctly.

The honest part

Will AI engines read it?

Nobody can promise that today, and we will not. No major engine has confirmed that it consumes llms.txt. OpenAI, Google and Anthropic all document their crawlers, and none of those documents mentions the file. Anyone selling llms.txt as a ranking tactic is ahead of the evidence.

Here is the case for serving one anyway. The file costs nothing to host, takes minutes to make, and cannot hurt: it is a plain text file no browser ever loads. And it is ready the day any engine adopts the convention, which is when prepared sites benefit first. Cheap preparation with no downside is worth 10 minutes. That is the whole pitch, and it is enough.

Beyond the file

The file is the easy part.

An llms.txt curates your pages. Whether those pages earn a recommendation is a different question, and that is the part Cruelx measures: a free report scores your site on design, SEO, copywriting, trust signals and buyer psychology, with fixes for what it finds.

Questions

llms.txt FAQ

What is an llms.txt file?

llms.txt is a plain text file served at a website’s root that gives AI assistants a curated index of the site: the site’s name as an H1, a 1-sentence summary in a blockquote, and short markdown lists of the pages that matter most. It was proposed at llmstxt.org in 2024 as a convention, the way robots.txt and sitemap.xml began.

Where do I put the llms.txt file?

At the root of your site, so it answers at yoursite.com/llms.txt, served as plain text. The format also allows a file on a subpath, like docs.yoursite.com/docs/llms.txt, and a reader is expected to use the most specific file that covers a URL. The root file is the one to start with.

Does llms.txt improve rankings or AI visibility?

There is no evidence it does, and this page will not pretend otherwise. No major engine has confirmed that it consumes llms.txt today. The case for serving one is different: it costs nothing, takes minutes, cannot hurt, and if the engines adopt the convention, prepared sites benefit first. Treat it as cheap preparation, not as a ranking tactic.

Can llms.txt block AI crawlers the way robots.txt does?

No. The 2 files answer opposite questions. robots.txt tells crawlers what they may fetch: it grants and denies access. llms.txt tells a reader what is worth fetching: it curates. A site can block an AI crawler in robots.txt and serve a perfect llms.txt at the same time, and the block wins, because permission comes before curation.

What goes in the Optional section?

Secondary links: pages a reader can skip when it needs a shorter context. By convention everything under an H2 heading named Optional is treated as skippable, and everything above it as the core index. If you are unsure, leave the section out. The generator does not add one, because it cannot know which of your pages are secondary.

Is the llms.txt generator free?

Yes. No signup, no email and no limit for normal use. The draft is composed from your site’s own pages with no AI call involved, which costs us nothing, so there is nothing to charge for.

What else can Cruelx tell me about my website?

Cruelx makes 2 products: an AI-powered website audit and the AI Visibility Scan, which measures whether AI answer engines recommend your business. The free website report scores your site on design, SEO, copywriting, trust signals and buyer psychology, then adds the measured checks: security grade, SSL certificate, DNS and email trust, robots and sitemap, and detected technologies. It takes under a minute and needs no card. See a sample report.

The full picture

Then measure it for real.

The AI Visibility Scan asks 270 questions across all 6 surfaces, keeps every answer as evidence, and ships a prioritized fix list. One time, no subscription.